feat: support multiple job IDs in roborev wait

[mariusvniekerk]

Summary

• Extend roborev wait to accept multiple job IDs/refs, waiting for all concurrently
• roborev wait --job 251 252 253 or roborev wait 251 252 253 now works
• Exit code 1 if any job fails; 0 only if all pass
• Goroutines poll in quiet mode to avoid interleaved output

Test plan

• TestWaitMultipleJobIDs — all jobs pass, expect exit 0
• TestWaitMultipleJobIDsOneFails — one fails, expect exit 1
• TestWaitMultipleJobIDsValidation — --sha incompatible with multiple args
• All 52 existing wait tests still pass
• Full suite (3503 tests) passes

🤖 Generated with Claude Code

[roborev-ci]

roborev: Combined Review (997c637)

Verdict: The PR introduces multi-job
wait functionality with a few bugs regarding error handling and missing output, but no security vulnerabilities were found.

🔴 High

• Silent failure path / Missing review output in multi-job mode
  • File: cmd/roborev/wait.go (lines 151-1
    63, 220, 244)
  • Description: The implementation calls waitForJob with quiet=true, which suppresses review output. While the code checks for errors, it never displays the actual review text or actionable output for job failures (such as *exitError). This
    silently drops results and can return exit code 1 in non-quiet mode without providing the user with any context.
  • Suggested Fix: After wg.Wait(), iterate over the completed jobs to fetch and print the actual review results serially, or capture the per-job outcome/output from
    waitForJob to display a summary later.

🟡 Medium

• First-error selection can mask a more important later error
  • File: cmd/roborev/wait.go (lines 235, 249)
  • Description:
    Error handling retains only the first error by argument order. If the first job returns ErrJobNotFound but a subsequent job experiences a real operational failure (e.g., transport/server error), the function may return exitError{1} and hide the actual root cause.
  • Suggested Fix: Prior
    itize returning non-*exitError/non-ErrJobNotFound errors first, or optionally aggregate errors and return the most actionable one.

---

Synthesized from 4 reviews (agents: codex, gemini | types: default, security)

[roborev-ci]

roborev: Combined Review (e313f91)

Verdict: All agents agree the code is clean; no medium, high, or critical issues were found.

---

Synthesized from 4 reviews (agents: codex, gemini | types: default, security)